This article is made on request of Miguel Neves from Skift Meetings.
The start of MeetMatch was to use AI for matchmaking. This was a very hard task, until Google and Facebook decided to release their AI code in open source around 2017. We then quickly assimilated that, to deliver the only business matchmaking solution within an event solution in the world.
In total, we use AI in:
- Profile analysis
We scan the Internet, looking for information about the companies: Website, Social media, Wikipedia page, Patent databases, … Our AI then processes this information to come to an extensive company profile. MeetMatch is the only event solution to do this. It is clear that this provides much better information than if you don’t do this.
We scan all information provided by the user, to come to an extensive participant profile.
- Processing extra inputs of attendees
MeetMatch supports different network formats. In several of them, attendees can e.g. indicate why they do/ do not want to meet a certain person, writing e.g. ‘We are already working with them’, ‘We are already in discussions with them’,’For this event, we are not interested in meeting Public Sector accounts’, … Our AI automatically understands these “commands” and translates them into the necessary matchmaking and scheduling actions. Overall, this information gives very good input on which meetings to optimally provide to attendees.
This is rather a nice consequence of our use of AI throughout the platform. This allows to cluster, perform allocations to domains, … and provide a great way to give visual insights into the event/ community.
This is the core of MeetMatch. The AI takes in all the inputs (profile, extra inputs, voting, past behavior, …) and defines the matching scores between each pair of participants. The implementation is actually a mix between AI and carefully selected algorithms, based on 8+ years of study of academic literature on business matchmaking.
- Provide innovation inspiration
Our matchmaking algorithm has 3 main modes: innovation, lead generation and job matching. In the innovation mode, the algorithm can provide different extra input on what to talk about: possible innovation projects, complementarities, …
- Filter spam
Many events are a bit assymetric: you like to have some attendees which attract other attendees, e.g. CEOs with a large budget. The flipside is that there is a risk that many attendees try to contact these “popular” attendees. To make sure the attractive attendees will keep visiting your event, it is good to protect them by avoiding spam.
- Face recognition
To be used for check-in at an event or session.